have a look at this article : https://appliednetsci.springeropen.com/track/pdf/10.1007/s41109-019-0165-9.pdf
this article aims to introduce CDLIB, describing its main features, and placing it among other tools already available to social network analysis practitioners.
Influence estimation ism however, somewhat tricky in Python. I particularly like the Key Player algorithm, Borgatti proposed herefor but since I did not come across a smooth implemetation of it in Python (here is an attempt for the positive version of KPP https://stats.stackexchange.com/questions/121645/borgatti-key-player-problem-kpp-in-python), I tend to switch over to UCINET for it.